Ever-increasing demand for data storage capacity has fostered the development of improved data array storage systems wherein a plurality of data storage devices is electronically linked to function synergistically. Data integrity schemes are also enhanced in such arrays permitting fail-safe redundant storage of data, such as in redundant arrays of independent device (“RAID”) systems.
There are a number of challenges facing the array designer. For example, the many and complex mechanical and electrical connections required for each data storage device are multiplied by the number in an array. That is, each and every data storage device requires sufficient mechanical support to isolate the delicate head and disc components from vibration levels that create data transfer errors. Not only must attention be paid to self-excitation, that is, vibration caused by the rotating disc of a data storage device itself, but like attention is required to external excitation sources in such an environment. External excitation can come from other data storage devices in the array, electrical components in the array such as power supplies and fans, and from the installation and/or removal of data storage devices while the array is operational.
As the number of data storage devices in arrays increases, the problems associated with electromagnetic interference containment are exacerbated as well. Properly shielding the data storage devices requires attention paid not only to leak paths between drives in adjacent shelves, but also to the leak paths potentially created by the multiple openings into which each of the plurality of data storage devices is inserted. Adequate shielding of these openings must be provided while still permitting the ability to insert and/or remove a data storage device without disrupting the shielding in place for adjacent data storage devices in the array.
Flexibility can be a problem as well. For instance, the electrical systems, such as the connector boards, controllers, and connecting buses, are hard-wired for a predetermined number and size of data storage devices in the array. This is required to maintain the electrical integrity of the array while permitting swapping of individual data storage devices. For this reason, the storage shelves and the associated electrical systems are dedicated for the predetermined number and size of data storage devices. Accordingly, because of both mechanical and electrical constraints, an array designed for a particular form factor configuration cannot readily be adapted for use with a different form factor. Also, if a grouping of data storage devices is needed for a particular function, such as mirroring the storage of data, such functionality must conventionally be achieved at the top level host programming level. This requires complex and coordinated programming of many data storage devices.
Another flexibility concern is control over the procedure for removing a data storage device, or a group of data storage devices, from service within the array. In use, each data storage device stores some data and commands in volatile memory. During a normal shutdown procedure all needed data and commands, as well as other needed information such as system state information, is written to nonvolatile memory for later use. Removing a data storage device from service before such a shutdown procedure is executed results in lost data and array system control errors.
While various approaches have been proposed in the art to address maximizing the data storage capacity while also providing operable flexibility in the utilization of data storage devices in array storage systems, there nevertheless remains a continued need for improvements in the art. It is to such improvements that the claimed invention is directed.